The Tests

India stated that the three nuclear tests, codenamed Shakti (Strength), were conducted on May 11, 1998 and the last two on May 13, 1998. The tests were conducted in different shafts, which had been dug in the previous two decades. Mrs. Indira Gandhi authorized the S-1 shaft [over 200m deep] and the S-2 shaft [over 150m deep], in 1982. The shafts were sized respectively for a boosted fission weapon, four times the size of POK-I and an advanced fission weapon of same size as POK-I [8]. Prime Minister Narasimha Rao authorized the third shaft in 1995 for a sub-kt test. PM Deve Gowda authorized the other two shafts in 1996 and PM Inder Gujral authorized another unused shaft. Clearly, the program had varying degrees of support from all the Prime Ministers in office despite differing political views. India came close to testing on three different occasions: 1982 under Mrs. Gandhi, 1995 under Mr. Rao and 1996 under Mr. Vajpayee. A brief description of the various tests and official Indian claims is in Table 1.

The sequence and location of these tests are critical in understanding how the measured data is analyzed. There are three types of analyses in the open literature. The primary analysis is based on the seismic measurements of the body wave magnitude of the seismic event itself. These values are correlated to the yields by means of correlations that are dependent on the site geology and subject to calibration accuracy. A second means of analyses is based on the radio chemical isotopes that are measured. These can indicate the extent and efficiency of the fission and fusion devices. A third means is by comparison of the surface topography after the test to other tests. This is also subject to the geologic characteristics of the site.

Western skepticism

Initial Western reaction concurred with the Indian claims. The USGS reported a seismic reading of 5.2 for mB on the Richter scale, which was similar to the Indian reports for S-1 and S-2. Roger Clark and Peter Zimmermann suggested agreement with the announced yields. A direct comparison of this measurement to the Pokhran-I test indicates that the measured yield exceeds the POK-I test by 2.4 to 2.5 times (7). The distinction between POK-I and POK-II is that the former test involved only one device, whereas the latter test involved two simultaneous tests and involves the concomitant wave interference effects.

Soon after, there were a number of counter claims (9) questioning the results of the tests. Wallace (10), and Barker (11) led these challenges. In summary, these concluded that the Indian tests did not provide results to match the announced values. The Federation of American Scientists (FAS) provides a comprehensive rebuttal of these early reports taking into account the recent work on the subject (7). The FAS study is closer to the Indian position. It, however, estimates the S-1 yield was only about 30kT, compared to the Indian claims of 43kT. This estimate is based on the relative magnitude of the seismic measurements of POK-1 and S-1/S-2, as indicated above.

Indian Response

The Bhabha Atomic Research Center (BARC) responded to the preliminary objections by publishing a series of papers arguing that the yields were in accordance with the design objectives. The first paper (13, 13 a) argues in favor of the necessity of accounting for the interference in seismic waves caused by the simultaneous S-1 and S-2 detonations. As the devices were aligned in an E-W direction, they determined, from physical considerations, that interference of the seismic waves of the twin explosions could only be neglected in a narrow azimuthal window. This window was determined to be +/- 20 degrees off the North South axis. They also updated the seismic constants used for mB wave magnitude for Pokhran site.

Their next paper (14) specifically addressed the Wallace and Barker papers and argued for the necessity of accounting for wave interference and site bias. In addition, they compared close-in ground accelerations from measurements to those values published in the scientific literature for similar events. A third approach taken was to compare the values to those of the earlier test in 1974. Based on all these three approaches BARC argued that the results were acceptable.

BARC followed up with another paper based on the regional waves, which are not detected by far away seismic stations (15). These waves are not detectable over tele-seismic distances and hence are unavailable to other stations. Here BARC compared the Lg & Rayleigh long period waves and again substantiated their case.

Discrepancy in Analyses

The determination of yield by seismic methods is not an easy process. Seismic records of previous tests do not sufficiently characterize the test site at Pokhran, nor is the site earthquake prone to calibrate the site. There are also site bias errors due to propagation of the short period seismic waves i.e. P-waves, which are detected from afar. A brief non-technical description of the seismic phenomena involved in the detection of nuclear explosions is provided here.

The basic process involved is the propagation of elastic waves in layered media, following the detonation of the explosive device. The explosive event results in the generation of waves-long period Rayleigh waves with an attendant Airy phase and the short period P - waves. The latter travel faster and are the first to arrive at distant locations, which are also called tele-seismic distances. The Rayleigh waves attenuate as an inverse function of distance and are limited to regional observations. Consequently, Rayleigh waves are not easily measured at great distances from the seismic event. The Rayleigh waves, are the first surface waves to arrive at regional locations and these are followed by the Lg phase of very fast oscillations of increasing amplitude. Following this, there is a big jump in amplitude due to the Airy phase which start diminishing in amplitude (Lg phase). However, the Airy phase diminishes inversely as a function of the distance raised to the five-sixths power. This is a simple consequence of the stationary phase principle for oscillatory integrals with degenerate phase (31). In other words, the Airy phase attenuates more slowly than the regular Rayleigh waves and accounts for the jump in amplitude in the seismograms.

Wallace’s methodology only addresses the P-waves that travel faster and have shorter wavelength. They are thus affected by local anomalies more than the Rayleigh waves, which have longer wavelength. If one looks at seismograms of the POK-2 from stations outside India these, record only the P-waves whereas those from India record all the local wave phenomena. However, to the casual observer it appears that the Nilore station should have picked up the regional seismic wave patterns, as it is located in the Rayleigh wave area. The fact those seismograms in India do show these patterns indicate that there is a seismic barrier between the Indian plate and the location of Nilore. It is likely that the Himalayan fault is responsible for the non-propagation of the surface waves to Nilore and the Kyrgyz network. This could account for the discrepancies in the seismograms in India and outside. Clearly, the complete seismic signature of the POK-II event was not detected outside India.

Table 2 indicates that the optimum viewing angle for seismic waves greater than 2.5 Hz is +/- 20 degrees of the N-S axis passing through the test site for minimal distortion. One has to ascertain the frequency content of the Indian explosions. This is graphed in Figure 6 of reference 26. The POK-II explosion had a large component in the high frequencies. In fact, Gupta et al in (26) and (20) specifically state that the seismic signals from the POK-II event peaked in the frequency range 3.5 to 6 Hz. Hence as is evident from Table 2, the method (adopted by BARC) of selecting mB values in certain limited stations along the N-S axis reduces the error due to interference. Since the Gauribidanur Seismic Array Station (GBA) is roughly 26 degrees off the North- South axis through Pokhran, reference to Table 2 shows that seismic waves from the POK-II event underwent destructive interference at the GBA location. Thus an upward correction to the mB value is warranted. As a result, of this interference related correction, the combined yield of the devices is estimated to have a body wave magnitude of 5.4. This corresponds to a yield around 5 times that of POK-I (7). Sublette indicates some reluctance to admit the corrections to the GBA data, but does not address the scientific aspects of wave interference. Based on Indian and recent Western claims regarding POK-I, it is reasonable to conclude that S-1/S-2 had a combined yield of around 50-55kT.

A study of Figure 2 in reference 15 shows the various components in the seismic wave resulting from a test explosion. It typically comprises of the P, Pn, Lg, Rayleigh wave with Airy phase. The figure documents the seismogram for the POK-II event as recorded at Bhopal (Figure 2a) and Nilore (Figure 2b). Remarkably, measurements at Nilore do not show the surface waves or Rayleigh waves. It only shows the faster moving short period waves that are affected by local anomalies. These components are not present in the seismograms outside India as they attenuate and do not cross over seismic faults. A study of the seismograms indicates similar anomalies in seismograms of Chagai at Bhuj, and Ajmer (26).

The anomaly in seismic signature was noted as early as the May 17, 1998 by Negi (34) who noted that the POK-I had a higher Richter reading than POK-II. The lower reading from a larger explosion can be explained by the interference from the simultaneous initiation of S-1 and S-2.

BARC published their radio-chemical analysis estimate of the S-1 yield (17). The raw data has not been presented as it could reveal the specifics of the weapon design. However, it provides a qualitative method of determining the efficacy of the tests. The paper also reveals the various decay elements found in the post-shot debris. Among the elements detected are isotopes of Ruthenium, Cesium, Manganese and Scandium. However, the analysis was done based only on those isotopes, which do not decay further e.g. 54 Mn and 46 Sc. It is instructive to compare the decay products detected in S-1 debris and those depicted in a paper by Drell (22). It could provide insight into the design used and indicates conclusively that fusion had occurred.

A preliminary estimate from radio-chemical data for S-2 is given as 13 +/- 3 kt (16). The radio-chemical confirmation of the yields of S-3, S-4 & S- 5 is also provided (18). This analysis also does not have raw data but gives a graph of the fission bye-products detected in the test holes and shows reaction did occur.

The radio-chemical analysis also provides the radius of cavity for the S-1 device and states it as 40 +/- 4m. Based on Trehune (23) and the knowledge that the test occurred in granite it is possible to independently confirm the yield. This method of estimation results in a nominal yield of 37kT. This compares favorably with the Indian report of 43 +/- 3 kT.

The basis of the radius of cavity estimate is not clarified. In the same paragraph, there are statements about the crush zone radius and the distance of the test hole to verify debris composition. In the error analysis section it is stated that the uncertainty in the estimate of the radius of the cavity is within 15 percent. Lamb (33) states that the yield estimate based on the radius of cavity determined by CORRTEX technique is within 15 percent of that determined by radio-chemical analysis.

While it is not clear in all probability BARC determined the radius of the cavity by means of this above technique which is arguably the most accurate method. Carey Sublette (12) confirms that this method was used during the tests.

Crater evidence

Sublette has given an excellent overview of the crater evidence and has ably refuted Wallace’s claims on that subject. Based on the surface features, Sublette concluded that the S-2 estimate based on crater evidence was reasonably close to what was claimed. This further suggests that the yield of S-2 was close to 10-12 kT. If this was the case, it is reasonable to expect that S-1 had a yield of 38-45 kT, which is well within the range claimed by BARC.

In an attempt to independently confirm this yield from the crater morphology, Sublette suggests that the S-1 crater evidence indicates that it was either over-buried or an imperfect explosion. Given that POK-I was buried at 107m and had a yield of about 10 kt, it is possible to scale the required depth for a device four times as powerful in the same terrain. As the depth is proportional to the cube root of the yield, the scaled depth is about 170m in similar terrain. S-1 was reported in various sources to be buried in a shaft over 200m and in granite. Hence, it is reasonable to accept that the device was over-buried. Chengappa reports that Chidambaram and Sikka were aware of this aspect and had stated this before informing the Prime Minister of the successful tests (28)

Sublette also questions the sub-kiloton tests based on the lack of imagery and any independently reported seismic evidence. The seismograms for POK-II at Nilore show that it did not detect many of the components of the larger explosions. It is possible that it would be challenging to detect much smaller signatures at Nilore. Sublette is not convinced of the radio-chemical evidence presented by BARC (18), although he does not question their veracity. However, BARC insists that the tests occurred as reported (21)

Discussion

In a limited series of tests BARC scientists have, by their own account, tried out a dozen ideas using three basic designs - fission, boosted fission and fusion. As the program was not overt, they had to operate under severe constraints. Ideally, they should have proofed S-1 to its full yield. However, that would mean acquiring and evacuating the villages around the test site and would have disclosed their plans. Hence, they tested the basic design and reduced yield. This is similar to the US and USSR programs in later sixties and early seventies respectively (30). The tests confirm that the fission process in its variations is fully understood. This is critical, as the fission process is required to be most reliable aspect of a deterrent posture (30).

Chengappa (28) reports that reactor grade plutonium was used for one of the devices tested. It is unclear which of the S-3 – S-5 tests used the Reactor Grade Plutonium. Kakodkar (4) says that the test site could handle 60 kilotons and other reports suggest the total design yield was 55 kilotons. In all probability, the S-3 shaft was capable of the difference. Besides Chengappa (29) states that the decision to conduct simultaneous tests was taken just before the test date. The concern was that the S-1 explosion could damage the other shafts. Considering this, the S-3 was conducted and probably the gas boosting was turned off as it could lead to crossing the site threshold. Further information on the radio-chemical debris is required to determine conclusively if reactor grade material was used. The data presented shows only that Cs 137 was detected at various depths of the shaft.

The reactor grade material significantly increases the options for India. Considering the material available from the CANDU reactors, India has the potential to increase the number of weapons to handle any changes in the international security environment. Previously, most Indian experts had suggested a minimum inventory in the low hundreds based on weapons grade Plutonium. The decision to make use of reactor grade material is a far-reaching one and needs further clarification.

No discussion of the Indian tests is complete without reference to the Pakistani tests at Chagai in the same month. Much work has to be done to understand those tests. In light of the geology of the terrain, it is possible that the yield was greater than estimated based on short period waves or mB data. In fact, Sikka comments on this possibility (16). Further Chanillo (20) compares the relative yields of the two events and concludes that Chagai was at most half of the POK-II.

Conclusions

We can draw the following conlcusions based on the preceeding analysis:

• BARC has presented seismic estimates of the total yield based on short period P waves, long period surface waves, and close in accelerations. These estimates give the total yield for the May 11, 1998 tests as about 55kT. In addition, they have presented results of the radio-chemical analysis. These radio chemical estimates, give the S-1 yield as 50 +/- 10 kT. The secondary data, including the radius of the cavity provides an estimate of 37 kT. This is close to the minimum design value of 40 kT. The S-1 device performed to the design within expected margins. The radio-chemical data shows decay products from a fusion reaction.
• The combined yield of S-1/S-2 less the possible yield of S-1 (from the radius of cavity calculation) indicates that S-2 was around 10-12kT. This is independently validated by the imagery of the S-2 crater.

The authors wish to thank Sagun Chanillo and Sanjay Badri Maharaj for their deep support and clarifications. Without their encouragement and support this paper would not have been put together.

[1] Chengappa, R., "Weapons Of Peace" Harpers Collins, India , ISBN 81-7223-330-2 page 432-433. It is stated that S-2 was a weapon from the inventory and was one of the old stock. See also fas.org web site - http://www.fas.org/nuke/hew/India/IndiaShakti.html. For an objective summary of the tests from a Western expert read http://www.thebulletin.org/issues/1998/ja98/ja98albright.html

[2a] Badri- Maharaj, Sanjay, " Armageddon Factor", Lancers , 2000. Page 79 ISBN 81-7062-109-7

[2b] R. Ramachandran, " Pokhran-II The Technical Aspects" in Mattoo, Amitabh, ed " India’s Nuclear Deterrent - Pokhran II and Beyond", Har Anand, New Delhi 1998.

[3] Chengappa, WOP, page 332, page 382 for operational, and page 86 for setting up the study group for high-pressure physics

[4] Interview with Dr. Anil Kakodkar, "We got everything we wanted", Frontline, June 6, 1998 Vol 15, No. 12, http://www.the-hindu.com/fline/fl1512/15120290.htm

[5] Subramanian, T. S, " Technological success", Front line May 23- June 05, 1998. http://www.the-hindu.com/fline/fl1511/15110130.htmSee also 24.

[6] Hersh, Seymour, "The Samson Option" Random House, New York, page 179. Israel had crossed the threshold in early 1968. See also Aslam Beg’s article" What entails signing the CTBT?" The News, April 03, 2001. Pakistan had crossed the threshold in 1987 based on Gen. Aslam Beg’s insight.

[7] Chengappa, WOP page 34. Chidambaram asked authorization for six tests to validate a dozen ideas. Also see reference 21.

[8] Chengappa, WOP, page 246 for quote on shaft capability of four times POK-I. Shaft depth ‘close to 200m’ page 247.

[11] Barker, Brian et al. 1998. "Monitoring Nuclear Tests", Science, Vol. 281, 25 September 1998, pp. 1967-68.

[12] FAS web site update on India – "What are the real yields of India’s Test?" http://www.fas.org/nuke/hew/India/IndiaRealYields.html  

[13] S.K. Sikka, Falguni Roy, G.J. Nair. 1998. "Indian Explosions of 11 May 1998: An Analysis of Global Seismic Body Wave Magnitude Estimates", Current Science 75, no. 5, 10 September 1998, pg. 491. BARC collection of links on POK-II and rebuttal of Barker paper.http://www.barc.ernet.in/webpages/milestones/explosion.html

[14] S.K. Sikka, Falguni Roy, G.J. Nair, V.G. Kolvankar and Anil Kakodkar. 1998. "Update on the yield of May 11-13, 1998 Nuclear Detonations at Pokhran", BARC News Letter No. 178, Nov. 1998. Located at http://www.barc.ernet.in/webpages/letter/newsletter_year_1998/novletter.html

[17] S.B.Manohar, B.S.Tomar, S.S.Rattan, V.K.Shukla, V.V.Kulkarni and Anil Kakodkar. 1999. "Post Shot Radioactivity Measurements On Samples Extracted From Thermonuclear Test Site", BARC News Letter No. 186, July 1999. Located at http://www.barc.ernet.in/webpages/letter/newsletter_year_1999/july99.html#POST SHOT radioactivity measurements on SAMPLES.

[19] IDC data summarized at fas web site:http://www.fas.org/nuke/hew/India/TN5-11REB.txt

[20] Chanillo, Sagun, "Relative yields of Pokhran and Chaghai nuclear explosions" Current Science, vol. 79[Dec 25th, 2000], No.12, 1631-1632. http://64.4.52.7/cgi-bin/linkrd?_lang=EN&lah=27904f0777d27a5b672012e9fcfa43ab&lat=988902284&hm___action=http%3a%2f%2fwww%2eias%2eernet%2ein%2fcurrsci%2fdec252000%2f1629%2epdf

[22] Drell, Sidney, Hippel, Frank von, " Limited Nuclear War", Scientific American, November 1976. The accompanying chart shows the chain reaction of high energy neutrons from D-T fusion on a U-238 material. It identifies similar decay products. Further investigation by those familiar with radio chemistry would be necessary to see if there was a thin liner or yield enhancing material used in S-1.

[23] Terhune, Glenn. Et al, " Numerical Simulation of the Baneberry event", Nuclear Technology, Vol. 46, Nov., 1976. Page 159- 169 Gives the relationship for radius of cavity is Rc = k*[Y]^1/3 - k varies from 12 to 16. The lower value being for hard rock like granite. The paper has numerical calculations to show that this holds true. Using this relationship and knowing that the POK-1 cavity radius was around 30m [Ref. 14] and the information that it was not in granite gives a method to the yield for that explosion. Using k equals 14 leads to a yield of ~ 9.8kt for POK-I.

[24] Chowdri, Satybrata " What did Pokhran reveal?", Daily Excelsior; May 09, 1999 . This article discusses the various technologies used and also speculates on the purpose of the sub-kt tests.

[26] Harsh Gupta, S.N.Bhattacharya, M. Ravi Kumar, D. Sarkar, "Spectral characteristics of Pokhran and Chagai nuclear explosions". Current Science, vol 76[1999], 1117-1121 http://tejas.serc.iisc.ernet.in/currsci/april25/articles20.htm

[27] Chengappa, WOP, pages page 431- 432 on S-1 shaft depth and geological strata.

[28] Chengappa, WOP, Pages 417- 418 on reactor grade material used in one of the tests.

[29] Chengappa, WOP, pages 425 on last minute decision to explode simultaneously to ensure no fratricide from S-1.

[30] Evernden, Marsh, "Yields of US and Soviet Nuclear tests", Physics Today, August 1987. This paper documents the difficulties in using seismic estimates without correction for site bias. It further quotes US officials like Richard Wagner , former Asst to Secy of Defense, who state that testing the primary, is critical to the design of the weapon. He further suggests that the primary and an altered secondary could be tested which do not exceed the threshold values. It also states that from 1965-74 the US and the USSR tested their advanced primaries and very few secondaries.

[31] Ewing, W.M. , Jardetzky, W.S, and Press F. , Elastic Waves in Layered Media, McGraw-Hill [1957]